move libmpeg3 to multimedia/

13 files changed, 643 insertions, 0 deletions

diff --git a/multimedia/libmpeg3/DESCR b/multimedia/libmpeg3/DESCR
new file mode 100644
index 00000000000..ca31a820815
--- /dev/null
+++ b/multimedia/libmpeg3/DESCR
@@ -0,0 +1,12 @@
+Libmpeg3 decodes
+                    MPEG-1 Layer II Audio
+                    MPEG-1 Layer III Audio
+                    MPEG-2 Layer III Audio
+                    MPEG-1 program streams
+                    MPEG-2 program streams
+                    MPEG-2 transport streams
+                    AC3 Audio
+                    MPEG-2 Video
+                    MPEG-1 Video
+                    IFO files
+                    VOB files
diff --git a/multimedia/libmpeg3/Makefile b/multimedia/libmpeg3/Makefile
new file mode 100644
index 00000000000..b5cc6eb3bdb
--- /dev/null
+++ b/multimedia/libmpeg3/Makefile
@@ -0,0 +1,64 @@
+# $NetBSD: Makefile,v 1.1 2004/06/26 18:00:21 grant Exp $
+#
+
+DISTNAME=	libmpeg3-1.3
+PKGREVISION=	4
+CATEGORIES=	devel audio graphics multimedia
+MASTER_SITES=	http://heroines.sourceforge.net/
+
+MAINTAINER=	tech-pkg@NetBSD.org
+HOMEPAGE=	http://heroines.sourceforge.net/libmpeg3.php3
+COMMENT=	MPEG decoding library
+
+.if (${MACHINE_ARCH} == "i386")
+BUILD_DEPENDS+=	nasm-[0-9]*:../../devel/nasm
+.else
+CONFIGURE_ARGS+= --no-mmx --no-css
+.endif
+
+USE_BUILDLINK3= yes
+HAS_CONFIGURE=	yes
+USE_GNU_TOOLS+=	make
+
+ONLY_FOR_COMPILER=	gcc
+
+PTHREAD_OPTS+=	require
+
+post-configure:
+	${ECHO} "CFLAGS+=-O2" >>${WRKSRC}/global_config
+	${ECHO} "CFLAGS+=-I${BUILDLINK_DIR}/include" >>${WRKSRC}/global_config
+	${ECHO} "CFLAGS+=-D_FILE_OFFSET_BITS=64" >>${WRKSRC}/global_config
+
+INSTHDRS=	bitstream.h libmpeg3.h mpeg3atrack.h mpeg3css.h \
+		mpeg3demux.h mpeg3io.h mpeg3private.h \
+		mpeg3private.inc mpeg3title.h mpeg3vtrack.h timecode.h
+AINSTHDRS=	ac3.h mpeg3audio.h
+VINSTHDRS=	idct.h mpeg3video.h slice.h
+INSTBINS=	mpeg3cat mpeg3toc mpeg3split
+
+do-install:
+	${INSTALL_PROGRAM} ${WRKSRC}/work/dump ${PREFIX}/bin/mpeg3dump
+	for f in ${INSTBINS}; do \
+		${INSTALL_PROGRAM} ${WRKSRC}/work/$$f ${PREFIX}/bin; \
+	done
+	${INSTALL_DATA_DIR} ${PREFIX}/include/mpeg3
+	for f in ${INSTHDRS}; do \
+		${INSTALL_DATA} ${WRKSRC}/$$f ${PREFIX}/include/mpeg3; \
+	done
+	${INSTALL_DATA_DIR} ${PREFIX}/include/mpeg3/audio
+	for f in ${AINSTHDRS}; do \
+		${INSTALL_DATA} ${WRKSRC}/audio/$$f \
+			${PREFIX}/include/mpeg3/audio; \
+	done
+	${INSTALL_DATA_DIR} ${PREFIX}/include/mpeg3/video
+	for f in ${VINSTHDRS}; do \
+		${INSTALL_DATA} ${WRKSRC}/video/$$f \
+			${PREFIX}/include/mpeg3/video; \
+	done
+	${INSTALL_DATA} ${WRKSRC}/work/libmpeg3.a ${PREFIX}/lib
+	${INSTALL_DATA_DIR} ${PREFIX}/share/doc/html/libmpeg3
+	${INSTALL_DATA} ${WRKSRC}/docs/index.html \
+		${PREFIX}/share/doc/html/libmpeg3
+
+.include "../../mk/pthread.buildlink3.mk"
+.include "../../mk/bsd.pkg.mk"
diff --git a/multimedia/libmpeg3/PLIST b/multimedia/libmpeg3/PLIST
new file mode 100644
index 00000000000..99fa16ab8f6
--- /dev/null
+++ b/multimedia/libmpeg3/PLIST
@@ -0,0 +1,27 @@
+@comment $NetBSD: PLIST,v 1.1 2004/06/26 18:00:21 grant Exp $
+include/mpeg3/libmpeg3.h
+bin/mpeg3dump
+bin/mpeg3cat
+bin/mpeg3toc
+bin/mpeg3split
+include/mpeg3/bitstream.h
+include/mpeg3/mpeg3atrack.h
+include/mpeg3/mpeg3css.h
+include/mpeg3/mpeg3demux.h
+include/mpeg3/mpeg3io.h
+include/mpeg3/mpeg3private.h
+include/mpeg3/mpeg3private.inc
+include/mpeg3/mpeg3title.h
+include/mpeg3/mpeg3vtrack.h
+include/mpeg3/timecode.h
+include/mpeg3/audio/ac3.h
+include/mpeg3/audio/mpeg3audio.h
+include/mpeg3/video/idct.h
+include/mpeg3/video/mpeg3video.h
+include/mpeg3/video/slice.h
+lib/libmpeg3.a
+share/doc/html/libmpeg3/index.html
+@dirrm include/mpeg3/audio
+@dirrm include/mpeg3/video
+@dirrm include/mpeg3
+@dirrm share/doc/html/libmpeg3
diff --git a/multimedia/libmpeg3/buildlink3.mk b/multimedia/libmpeg3/buildlink3.mk
new file mode 100644
index 00000000000..10aaf508dd5
--- /dev/null
+++ b/multimedia/libmpeg3/buildlink3.mk
@@ -0,0 +1,20 @@
+# $NetBSD: buildlink3.mk,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+BUILDLINK_DEPTH:=		${BUILDLINK_DEPTH}+
+LIBMPEG3_BUILDLINK3_MK:=	${LIBMPEG3_BUILDLINK3_MK}+
+
+.if !empty(BUILDLINK_DEPTH:M+)
+BUILDLINK_DEPENDS+=	libmpeg3
+.endif
+
+BUILDLINK_PACKAGES:=	${BUILDLINK_PACKAGES:Nlibmpeg3}
+BUILDLINK_PACKAGES+=	libmpeg3
+
+.if !empty(LIBMPEG3_BUILDLINK3_MK:M+)
+BUILDLINK_DEPENDS.libmpeg3+=	libmpeg3>=1.3nb4
+BUILDLINK_PKGSRCDIR.libmpeg3?=	../../multimedia/libmpeg3
+.endif	# LIBMPEG3_BUILDLINK3_MK
+
+.include "../../mk/pthread.buildlink3.mk"
+
+BUILDLINK_DEPTH:=     ${BUILDLINK_DEPTH:S/+$//}
diff --git a/multimedia/libmpeg3/distinfo b/multimedia/libmpeg3/distinfo
new file mode 100644
index 00000000000..cbf0204a06a
--- /dev/null
+++ b/multimedia/libmpeg3/distinfo
@@ -0,0 +1,12 @@
+$NetBSD: distinfo,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+SHA1 (libmpeg3-1.3.tar.gz) = e7e8b03ab45d5593558b0d19e31f0694cfdfa2b5
+Size (libmpeg3-1.3.tar.gz) = 154062 bytes
+SHA1 (patch-aa) = a5665a330234b964aca00a5c4c6f00a9f1bd11ff
+SHA1 (patch-ab) = 844cfcbd99269f1f0b4bef76e0a8567522070707
+SHA1 (patch-ac) = fa32047a70936d0ab81100a4f4b3b0ccdcd3a708
+SHA1 (patch-ad) = 86edf9af264370bf89050b02e956b5bacc86881e
+SHA1 (patch-ae) = 55cf13476c59da1844fb4bc2bff7b50a78a6a4f1
+SHA1 (patch-af) = 7e62a236ee0a7398c1b1a3e994261ebb9be50bae
+SHA1 (patch-ag) = bf321ac689b82889ca1a2bb8a96cb6e387280f32
+SHA1 (patch-ah) = 58ffda3df9be3b089de362451ddf249a74fac5be
diff --git a/multimedia/libmpeg3/patches/patch-aa b/multimedia/libmpeg3/patches/patch-aa
new file mode 100644
index 00000000000..d697c21a9e6
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-aa
@@ -0,0 +1,22 @@
+$NetBSD: patch-aa,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- Makefile.orig	Sun May 20 05:05:26 2001
++++ Makefile	Tue Jul 31 17:34:17 2001
+@@ -2,7 +2,7 @@
+ 
+ CFLAGS += 
+ 
+-OBJDIR = $(shell uname --machine)
++OBJDIR = work
+ 
+ $(shell sh -c 'if ! test -d $(OBJDIR)\; then \
+ 	mkdir $(OBJDIR)\; \
+@@ -69,7 +69,7 @@
+ 
+ OUTPUT = $(OBJDIR)/libmpeg3.a
+ UTILS = $(OBJDIR)/dump $(OBJDIR)/mpeg3cat $(OBJDIR)/mpeg3toc $(OBJDIR)/mpeg3split
+-LIBS = -lm -lpthread
++LIBS = -L${LOCALBASE}/lib -Wl,-R${LOCALBASE}/lib -lm -lpthread
+ 
+ all: $(OUTPUT) util
+ 
diff --git a/multimedia/libmpeg3/patches/patch-ab b/multimedia/libmpeg3/patches/patch-ab
new file mode 100644
index 00000000000..1a3e8b09b94
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-ab
@@ -0,0 +1,13 @@
+$NetBSD: patch-ab,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- mpeg3css.c.orig	Tue Jul 31 17:16:17 2001
++++ mpeg3css.c	Tue Jul 31 17:16:44 2001
+@@ -45,7 +45,7 @@
+ #include <sys/types.h>
+ #include <sys/stat.h>
+ #include <sys/types.h>
+-#include <linux/cdrom.h>
++#include <sys/dvdio.h>
+ 
+ #ifndef FIBMAP
+ #define FIBMAP	   _IO(0x00,1)	/* bmap access */
diff --git a/multimedia/libmpeg3/patches/patch-ac b/multimedia/libmpeg3/patches/patch-ac
new file mode 100644
index 00000000000..987729d8d3a
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-ac
@@ -0,0 +1,19 @@
+$NetBSD: patch-ac,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- mpeg3ifo.c.orig	Tue Jul 31 17:18:25 2001
++++ mpeg3ifo.c	Tue Jul 31 17:18:56 2001
+@@ -1,9 +1,13 @@
+-#include <byteswap.h>
+ #include <dirent.h>
+ #include <fcntl.h>
+ #include <stdlib.h>
+ #include <string.h>
+ #include <sys/types.h>
++
++#include <sys/endian.h>
++#include <machine/bswap.h>
++#define bswap_16(x) bswap16(x)
++#define bswap_32(x) bswap32(x)
+ 
+ #include "ifo.h"
+ #include "mpeg3private.h"
diff --git a/multimedia/libmpeg3/patches/patch-ad b/multimedia/libmpeg3/patches/patch-ad
new file mode 100644
index 00000000000..549eb529364
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-ad
@@ -0,0 +1,22 @@
+$NetBSD: patch-ad,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- ifo.h.orig	Tue Jul 31 17:19:54 2001
++++ ifo.h	Tue Jul 31 17:20:47 2001
+@@ -193,7 +193,7 @@
+ 	u_char *data[10];
+ 	
+ 	int fd;		// file descriptor
+-	__off64_t pos;	// offset of ifo file on device 
++	off_t pos;	// offset of ifo file on device 
+ } ifo_t;
+ 
+ 
+@@ -221,7 +221,7 @@
+  * Prototypes
+  */
+ 
+-ifo_t *ifoOpen (int fd, __off64_t pos);
++ifo_t *ifoOpen (int fd, off_t pos);
+ int ifoClose (ifo_t *ifo);
+ 
+ u_int ifoGetVOBStart	(ifo_t *ifo);
diff --git a/multimedia/libmpeg3/patches/patch-ae b/multimedia/libmpeg3/patches/patch-ae
new file mode 100644
index 00000000000..945eb272c37
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-ae
@@ -0,0 +1,55 @@
+$NetBSD: patch-ae,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- mpeg3io.c.orig	2001-05-20 13:05:26.000000000 +1000
++++ mpeg3io.c	2004-05-04 23:48:36.000000000 +1000
+@@ -1,11 +1,17 @@
+ #include "mpeg3private.h"
+ #include "mpeg3protos.h"
+ 
+-#include <mntent.h>
+-#include <sys/stat.h>
++#include <sys/param.h>
++#include <sys/mount.h>
+ #include <stdlib.h>
+ #include <string.h>
+ 
++#ifdef __NetBSD__
++# if __NetBSD_Version__ >= 200040000	/* 2.0D */
++#  define statfs statvfs
++# endif
++#endif
++
+ mpeg3_fs_t* mpeg3_new_fs(char *path)
+ {
+ 	mpeg3_fs_t *fs = calloc(1, sizeof(mpeg3_fs_t));
+@@ -88,27 +94,15 @@
+ 
+ int mpeg3io_device(char *path, char *device)
+ {
+-	struct stat file_st, device_st;
+-    struct mntent *mnt;
+-	FILE *fp;
++	struct statfs file_st;
+ 
+-	if(stat(path, &file_st) < 0)
++	if(statfs(path, &file_st) < 0)
+ 	{
+ 		perror("mpeg3io_device");
+ 		return 1;
+ 	}
+ 
+-	fp = setmntent(MOUNTED, "r");
+-    while(fp && (mnt = getmntent(fp)))
+-	{
+-		if(stat(mnt->mnt_fsname, &device_st) < 0) continue;
+-		if(device_st.st_rdev == file_st.st_dev)
+-		{
+-			strncpy(device, mnt->mnt_fsname, MPEG3_STRLEN);
+-			break;
+-		}
+-	}
+-	endmntent(fp);
++	strncpy(device, file_st.f_mntfromname, MPEG3_STRLEN);
+ 
+ 	return 0;
+ }
diff --git a/multimedia/libmpeg3/patches/patch-af b/multimedia/libmpeg3/patches/patch-af
new file mode 100644
index 00000000000..e890d31e5a9
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-af
@@ -0,0 +1,22 @@
+$NetBSD: patch-af,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- video/reconstruct.c.orig	Tue Jul 31 17:23:42 2001
++++ video/reconstruct.c	Tue Jul 31 17:24:52 2001
+@@ -572,7 +572,7 @@
+ 	}
+ }
+ 
+-static inline void recv(unsigned char *s, unsigned char *d, int lx, int lx2, int h)
++static inline void xxrecv(unsigned char *s, unsigned char *d, int lx, int lx2, int h)
+ {
+ 	unsigned char *dp,*sp,*sp2;
+ 	int j;
+@@ -971,7 +971,7 @@
+ 			case 0x0:	recc(s, d, lx2, h);       break;
+ 			case 0x7:   recva(s, d, lx, lx2, h);  break;
+ 			case 0x6:   recvac(s, d, lx, lx2, h); break;
+-			case 0x5:	recv(s, d, lx, lx2, h);   break;
++			case 0x5:	xxrecv(s, d, lx, lx2, h);   break;
+ 			case 0x4:	recvc(s, d, lx, lx2, h);  break;
+ 			case 0x9:	rech(s, d, lx2, h);       break;
+ 			case 0x8:   rechc(s, d, lx2, h);      break;
diff --git a/multimedia/libmpeg3/patches/patch-ag b/multimedia/libmpeg3/patches/patch-ag
new file mode 100644
index 00000000000..2a0f30fefb0
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-ag
@@ -0,0 +1,45 @@
+$NetBSD: patch-ag,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- video/mmxtest.c.orig	Tue Jul 31 19:37:07 2001
++++ video/mmxtest.c	Tue Jul 31 19:41:16 2001
+@@ -5,39 +5,9 @@
+ 
+ int mpeg3_mmx_test()
+ {
+-	int result = 0;
+-	FILE *proc;
+-	char string[MPEG3_STRLEN];
+-
+ 
+ #ifdef HAVE_MMX
+-	if(!(proc = fopen(MPEG3_PROC_CPUINFO, "r")))
+-	{
+-		fprintf(stderr, "mpeg3_mmx_test: failed to open /proc/cpuinfo\n");
+-		return 0;
+-	}
+-	
+-	while(!feof(proc))
+-	{
+-		fgets(string, MPEG3_STRLEN, proc);
+-/* Got the flags line */
+-		if(!strncasecmp(string, "flags", 5))
+-		{
+-			char *needle;
+-			needle = strstr(string, "mmx");
+-			if(!needle)
+-            {
+-            	fclose(proc);
+-            	return 0;
+-            }
+-			if(!strncasecmp(needle, "mmx", 3))
+-            {
+-            	fclose(proc);
+-            	return 1;
+-            }
+-		}
+-	}
+-   	fclose(proc);
++	return (1);
+ #endif
+ 
+ 	return 0;
diff --git a/multimedia/libmpeg3/patches/patch-ah b/multimedia/libmpeg3/patches/patch-ah
new file mode 100644
index 00000000000..12035bb0083
--- /dev/null
+++ b/multimedia/libmpeg3/patches/patch-ah
@@ -0,0 +1,310 @@
+$NetBSD: patch-ah,v 1.1 2004/06/26 18:00:21 grant Exp $
+
+--- video/output.c.orig	2001-05-20 13:05:26.000000000 +1000
++++ video/output.c	2003-07-29 00:53:45.000000000 +1000
+@@ -207,50 +207,50 @@
+ 		);
+ }
+ 
+-static unsigned long long  mpeg3_MMX_U_80 = 0x0000008000800000;
+-static unsigned long long  mpeg3_MMX_V_80 = 0x0000000000800080;
+-static long long  mpeg3_MMX_U_COEF        = 0x00000058ffd30000;
+-static long long  mpeg3_MMX_V_COEF        = 0x00000000ffea006f;
+-static long long  mpeg3_MMX_601_Y_COEF    = 0x0000004800480048;
+-static long long  mpeg3_MMX_601_Y_DIFF    = 0x0000000000000010;
++static unsigned long long  mpeg3_MMX_U_80 = 0x0000008000800000LL;
++static unsigned long long  mpeg3_MMX_V_80 = 0x0000000000800080LL;
++static long long  mpeg3_MMX_U_COEF        = 0x00000058ffd30000LL;
++static long long  mpeg3_MMX_V_COEF        = 0x00000000ffea006fLL;
++static long long  mpeg3_MMX_601_Y_COEF    = 0x0000004800480048LL;
++static long long  mpeg3_MMX_601_Y_DIFF    = 0x0000000000000010LL;
+ 
+ inline void mpeg3_bgra32_mmx(unsigned long y, 
+ 		unsigned long u, 
+ 		unsigned long v, 
+ 		unsigned long *output)
+ {
+-asm("
+-/* Output will be 0x00rrggbb with the 00 trailing so this can also be used */
+-/* for bgr24. */
+-	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */
+-	movd (%1), %%mm1;          /* Load u    0x00000000000000cr */
+-	movq %%mm0, %%mm3;         /* Copy y to temp */
+-	psllq $16, %%mm1;          /* Shift u   0x0000000000cr0000 */
+-	movd (%2), %%mm2;          /* Load v    0x00000000000000cb */
+-	psllq $16, %%mm3;          /* Shift y */
+-	movq %%mm1, %%mm4;         /* Copy u to temp */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */
+-	psllq $16, %%mm4;          /* Shift u */
+-	movq %%mm2, %%mm5;         /* Copy v to temp */
+-	psllq $16, %%mm3;          /* Shift y  */
+-	por %%mm4, %%mm1;          /* Overlay new u byte 0x000000cr00cr0000 */
+-	psllq $16, %%mm5;          /* Shift v  */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */
+-	por %%mm5, %%mm2;          /* Overlay new v byte 0x0000000000cb00cb */
+-
+-/* mm0: 0x000000yy00yy00yy mm1: 0x000000uu00uu0000 mm2: 0x0000000000vv00vv */
+- 	psubw mpeg3_MMX_U_80, %%mm1;    /* Subtract 128 from u 0x000000uu00uu0000 */
+- 	pmullw mpeg3_MMX_U_COEF, %%mm1; /* Multiply u coeffs 0x0000uuuuuuuu0000 */
+- 	psllw $6, %%mm0;                /* Shift y coeffs 0x0000yyy0yyy0yyy0 */
+- 	psubw mpeg3_MMX_V_80, %%mm2;    /* Subtract 128 from v 0x0000000000cb00cb */
+- 	pmullw mpeg3_MMX_V_COEF, %%mm2; /* Multiply v coeffs 0x0000crcrcrcrcrcr */
+-
+-/* mm0: 0x000000yy00yy00yy mm1: 0x0000uuuuuuuu0000 mm2: 0x00000000vvvvvvvv */
+-	paddsw %%mm1, %%mm0;        /* Add u to result */
+-	paddsw %%mm2, %%mm0;        /* Add v to result 0x0000rrrrggggbbbb */
+-	psraw $6, %%mm0;           /* Demote precision */
+-	packuswb %%mm0, %%mm0;     /* Pack into ARGB 0x0000000000rrggbb */
+-	movd %%mm0, (%3);          /* Store output */
++asm("\n\
++/* Output will be 0x00rrggbb with the 00 trailing so this can also be used */\n\
++/* for bgr24. */\n\
++	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */\n\
++	movd (%1), %%mm1;          /* Load u    0x00000000000000cr */\n\
++	movq %%mm0, %%mm3;         /* Copy y to temp */\n\
++	psllq $16, %%mm1;          /* Shift u   0x0000000000cr0000 */\n\
++	movd (%2), %%mm2;          /* Load v    0x00000000000000cb */\n\
++	psllq $16, %%mm3;          /* Shift y */\n\
++	movq %%mm1, %%mm4;         /* Copy u to temp */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */\n\
++	psllq $16, %%mm4;          /* Shift u */\n\
++	movq %%mm2, %%mm5;         /* Copy v to temp */\n\
++	psllq $16, %%mm3;          /* Shift y  */\n\
++	por %%mm4, %%mm1;          /* Overlay new u byte 0x000000cr00cr0000 */\n\
++	psllq $16, %%mm5;          /* Shift v  */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */\n\
++	por %%mm5, %%mm2;          /* Overlay new v byte 0x0000000000cb00cb */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy mm1: 0x000000uu00uu0000 mm2: 0x0000000000vv00vv */\n\
++ 	psubw mpeg3_MMX_U_80, %%mm1;    /* Subtract 128 from u 0x000000uu00uu0000 */\n\
++ 	pmullw mpeg3_MMX_U_COEF, %%mm1; /* Multiply u coeffs 0x0000uuuuuuuu0000 */\n\
++ 	psllw $6, %%mm0;                /* Shift y coeffs 0x0000yyy0yyy0yyy0 */\n\
++ 	psubw mpeg3_MMX_V_80, %%mm2;    /* Subtract 128 from v 0x0000000000cb00cb */\n\
++ 	pmullw mpeg3_MMX_V_COEF, %%mm2; /* Multiply v coeffs 0x0000crcrcrcrcrcr */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy mm1: 0x0000uuuuuuuu0000 mm2: 0x00000000vvvvvvvv */\n\
++	paddsw %%mm1, %%mm0;        /* Add u to result */\n\
++	paddsw %%mm2, %%mm0;        /* Add v to result 0x0000rrrrggggbbbb */\n\
++	psraw $6, %%mm0;           /* Demote precision */\n\
++	packuswb %%mm0, %%mm0;     /* Pack into ARGB 0x0000000000rrggbb */\n\
++	movd %%mm0, (%3);          /* Store output */\n\
+ 	"
+ :
+ : "r" (&y), "r" (&u), "r" (&v), "r" (output));
+@@ -261,39 +261,39 @@
+ 		unsigned long v, 
+ 		unsigned long *output)
+ {
+-asm("
+-/* Output will be 0x00rrggbb with the 00 trailing so this can also be used */
+-/* for bgr24. */
+-	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */
+-	psubsw mpeg3_MMX_601_Y_DIFF, %%mm0;      /* Subtract 16 from y */
+-	movd (%1), %%mm1;          /* Load u    0x00000000000000cr */
+-	movq %%mm0, %%mm3;         /* Copy y to temp */
+-	psllq $16, %%mm1;          /* Shift u   0x0000000000cr0000 */
+-	movd (%2), %%mm2;          /* Load v    0x00000000000000cb */
+-	psllq $16, %%mm3;          /* Shift y */
+-	movq %%mm1, %%mm4;         /* Copy u to temp */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */
+-	psllq $16, %%mm4;          /* Shift u */
+-	movq %%mm2, %%mm5;         /* Copy v to temp */
+-	psllq $16, %%mm3;          /* Shift y  */
+-	por %%mm4, %%mm1;          /* Overlay new u byte 0x000000cr00cr0000 */
+-	psllq $16, %%mm5;          /* Shift v  */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */
+-	por %%mm5, %%mm2;          /* Overlay new v byte 0x0000000000cb00cb */
+-
+-/* mm0: 0x000000yy00yy00yy mm1: 0x000000uu00uu0000 mm2: 0x0000000000vv00vv */
+-	pmullw mpeg3_MMX_601_Y_COEF, %%mm0; /* Scale and shift y coeffs */
+-	psubw mpeg3_MMX_U_80, %%mm1;     /* Subtract 128 from u 0x000000uu00uu0000 */
+- 	pmullw mpeg3_MMX_U_COEF, %%mm1;  /* Multiply u coeffs 0x0000uuuuuuuu0000 */
+-	psubw mpeg3_MMX_V_80, %%mm2;     /* Subtract 128 from v 0x0000000000cb00cb */
+- 	pmullw mpeg3_MMX_V_COEF, %%mm2;  /* Multiply v coeffs 0x0000crcrcrcrcrcr */
+-
+-/* mm0: 0x000000yy00yy00yy mm1: 0x0000uuuuuuuu0000 mm2: 0x00000000vvvvvvvv */
+-	paddsw %%mm1, %%mm0;        /* Add u to result */
+-	paddsw %%mm2, %%mm0;        /* Add v to result 0x0000rrrrggggbbbb */
+-	psraw $6, %%mm0;           /* Demote precision */
+-	packuswb %%mm0, %%mm0;     /* Pack into ARGB 0x0000000000rrggbb */
+-	movd %%mm0, (%3);          /* Store output */
++asm("\n\
++/* Output will be 0x00rrggbb with the 00 trailing so this can also be used */\n\
++/* for bgr24. */\n\
++	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */\n\
++	psubsw mpeg3_MMX_601_Y_DIFF, %%mm0;      /* Subtract 16 from y */\n\
++	movd (%1), %%mm1;          /* Load u    0x00000000000000cr */\n\
++	movq %%mm0, %%mm3;         /* Copy y to temp */\n\
++	psllq $16, %%mm1;          /* Shift u   0x0000000000cr0000 */\n\
++	movd (%2), %%mm2;          /* Load v    0x00000000000000cb */\n\
++	psllq $16, %%mm3;          /* Shift y */\n\
++	movq %%mm1, %%mm4;         /* Copy u to temp */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */\n\
++	psllq $16, %%mm4;          /* Shift u */\n\
++	movq %%mm2, %%mm5;         /* Copy v to temp */\n\
++	psllq $16, %%mm3;          /* Shift y  */\n\
++	por %%mm4, %%mm1;          /* Overlay new u byte 0x000000cr00cr0000 */\n\
++	psllq $16, %%mm5;          /* Shift v  */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */\n\
++	por %%mm5, %%mm2;          /* Overlay new v byte 0x0000000000cb00cb */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy mm1: 0x000000uu00uu0000 mm2: 0x0000000000vv00vv */\n\
++	pmullw mpeg3_MMX_601_Y_COEF, %%mm0; /* Scale and shift y coeffs */\n\
++	psubw mpeg3_MMX_U_80, %%mm1;     /* Subtract 128 from u 0x000000uu00uu0000 */\n\
++ 	pmullw mpeg3_MMX_U_COEF, %%mm1;  /* Multiply u coeffs 0x0000uuuuuuuu0000 */\n\
++	psubw mpeg3_MMX_V_80, %%mm2;     /* Subtract 128 from v 0x0000000000cb00cb */\n\
++ 	pmullw mpeg3_MMX_V_COEF, %%mm2;  /* Multiply v coeffs 0x0000crcrcrcrcrcr */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy mm1: 0x0000uuuuuuuu0000 mm2: 0x00000000vvvvvvvv */\n\
++	paddsw %%mm1, %%mm0;        /* Add u to result */\n\
++	paddsw %%mm2, %%mm0;        /* Add v to result 0x0000rrrrggggbbbb */\n\
++	psraw $6, %%mm0;           /* Demote precision */\n\
++	packuswb %%mm0, %%mm0;     /* Pack into ARGB 0x0000000000rrggbb */\n\
++	movd %%mm0, (%3);          /* Store output */\n\
+ 	"
+ :
+ : "r" (&y), "r" (&u), "r" (&v), "r" (output));
+@@ -309,38 +309,38 @@
+ 		unsigned long v, 
+ 		unsigned long *output)
+ {
+-asm("
+-/* Output will be 0x00bbggrr with the 00 trailing so this can also be used */
+-/* for rgb24. */
+-	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */
+-	movd (%1), %%mm1;          /* Load v    0x00000000000000vv */
+-	movq %%mm0, %%mm3;         /* Copy y to temp */
+-	psllq $16, %%mm1;          /* Shift v   0x0000000000vv0000 */
+-	movd (%2), %%mm2;          /* Load u    0x00000000000000uu */
+-	psllq $16, %%mm3;          /* Shift y */
+-	movq %%mm1, %%mm4;         /* Copy v to temp */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */
+-	psllq $16, %%mm4;          /* Shift v */
+-	movq %%mm2, %%mm5;         /* Copy u to temp */
+-	psllq $16, %%mm3;          /* Shift y  */
+-	por %%mm4, %%mm1;          /* Overlay new v byte 0x000000vv00vv0000 */
+-	psllq $16, %%mm5;          /* Shift u  */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */
+-	por %%mm5, %%mm2;          /* Overlay new u byte 0x0000000000uu00uu */
+-
+-/* mm0: 0x000000yy00yy00yy mm1: 0x000000vv00vv0000 mm2: 0x0000000000uu00uu */
+- 	psubw mpeg3_MMX_V_80_RGB, %%mm1;    /* Subtract 128 from v 0x000000vv00vv0000 */
+- 	pmullw mpeg3_MMX_V_COEF_RGB, %%mm1; /* Multiply v coeffs 0x0000vvvvvvvv0000 */
+- 	psllw $6, %%mm0;                /* Shift y coeffs 0x0000yyy0yyy0yyy0 */
+- 	psubw mpeg3_MMX_U_80_RGB, %%mm2;    /* Subtract 128 from u 0x0000000000uu00uu */
+- 	pmullw mpeg3_MMX_U_COEF_RGB, %%mm2; /* Multiply u coeffs 0x0000uuuuuuuuuuuu */
+-
+-/* mm0: 0x000000yy00yy00yy mm1: 0x0000vvvvvvvv0000 mm2: 0x00000000uuuuuuuu */
+-	paddsw %%mm1, %%mm0;        /* Add v to result */
+-	paddsw %%mm2, %%mm0;        /* Add u to result 0x0000bbbbggggrrrr */
+-	psraw $6, %%mm0;           /* Demote precision */
+-	packuswb %%mm0, %%mm0;     /* Pack into RGBA 0x0000000000bbggrr */
+-	movd %%mm0, (%3);          /* Store output */
++asm("\n\
++/* Output will be 0x00bbggrr with the 00 trailing so this can also be used */\n\
++/* for rgb24. */\n\
++	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */\n\
++	movd (%1), %%mm1;          /* Load v    0x00000000000000vv */\n\
++	movq %%mm0, %%mm3;         /* Copy y to temp */\n\
++	psllq $16, %%mm1;          /* Shift v   0x0000000000vv0000 */\n\
++	movd (%2), %%mm2;          /* Load u    0x00000000000000uu */\n\
++	psllq $16, %%mm3;          /* Shift y */\n\
++	movq %%mm1, %%mm4;         /* Copy v to temp */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */\n\
++	psllq $16, %%mm4;          /* Shift v */\n\
++	movq %%mm2, %%mm5;         /* Copy u to temp */\n\
++	psllq $16, %%mm3;          /* Shift y  */\n\
++	por %%mm4, %%mm1;          /* Overlay new v byte 0x000000vv00vv0000 */\n\
++	psllq $16, %%mm5;          /* Shift u  */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */\n\
++	por %%mm5, %%mm2;          /* Overlay new u byte 0x0000000000uu00uu */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy mm1: 0x000000vv00vv0000 mm2: 0x0000000000uu00uu */\n\
++ 	psubw mpeg3_MMX_V_80_RGB, %%mm1;    /* Subtract 128 from v 0x000000vv00vv0000 */\n\
++ 	pmullw mpeg3_MMX_V_COEF_RGB, %%mm1; /* Multiply v coeffs 0x0000vvvvvvvv0000 */\n\
++ 	psllw $6, %%mm0;                /* Shift y coeffs 0x0000yyy0yyy0yyy0 */\n\
++ 	psubw mpeg3_MMX_U_80_RGB, %%mm2;    /* Subtract 128 from u 0x0000000000uu00uu */\n\
++ 	pmullw mpeg3_MMX_U_COEF_RGB, %%mm2; /* Multiply u coeffs 0x0000uuuuuuuuuuuu */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy mm1: 0x0000vvvvvvvv0000 mm2: 0x00000000uuuuuuuu */\n\
++	paddsw %%mm1, %%mm0;        /* Add v to result */\n\
++	paddsw %%mm2, %%mm0;        /* Add u to result 0x0000bbbbggggrrrr */\n\
++	psraw $6, %%mm0;           /* Demote precision */\n\
++	packuswb %%mm0, %%mm0;     /* Pack into RGBA 0x0000000000bbggrr */\n\
++	movd %%mm0, (%3);          /* Store output */\n\
+ 	"
+ :
+ : "r" (&y), "r" (&v), "r" (&u), "r" (output));
+@@ -351,39 +351,39 @@
+ 		unsigned long v, 
+ 		unsigned long *output)
+ {
+-asm("
+-/* Output will be 0x00bbggrr with the 00 trailing so this can also be used */
+-/* for rgb24. */
+-	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */
+-	psubsw mpeg3_MMX_601_Y_DIFF, %%mm0;      /* Subtract 16 from y */
+-	movd (%1), %%mm1;          /* Load v    0x00000000000000vv */
+-	movq %%mm0, %%mm3;         /* Copy y to temp */
+-	psllq $16, %%mm1;          /* Shift v   0x0000000000vv0000 */
+-	movd (%2), %%mm2;          /* Load u    0x00000000000000uu */
+-	psllq $16, %%mm3;          /* Shift y */
+-	movq %%mm1, %%mm4;         /* Copy v to temp */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */
+-	psllq $16, %%mm4;          /* Shift v */
+-	movq %%mm2, %%mm5;         /* Copy u to temp */
+-	psllq $16, %%mm3;          /* Shift y  */
+-	por %%mm4, %%mm1;          /* Overlay new v byte 0x000000vv00vv0000 */
+-	psllq $16, %%mm5;          /* Shift u  */
+-	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */
+-	por %%mm5, %%mm2;          /* Overlay new u byte 0x0000000000uu00uu */
+-
+-/* mm0: 0x000000yy00yy00yy     mm1: 0x000000vv00vv0000     mm2: 0x0000000000uu00uu */
+-	pmullw mpeg3_MMX_601_Y_COEF, %%mm0;     /* Scale y coeffs */
+- 	psubw mpeg3_MMX_V_80_RGB, %%mm1;    /* Subtract 128 from v 0x000000vv00vv0000 */
+- 	pmullw mpeg3_MMX_V_COEF_RGB, %%mm1; /* Multiply v coeffs 0x0000vvvvvvvv0000 */
+- 	psubw mpeg3_MMX_U_80_RGB, %%mm2;    /* Subtract 128 from u 0x0000000000uu00uu */
+- 	pmullw mpeg3_MMX_U_COEF_RGB, %%mm2; /* Multiply u coeffs 0x0000uuuuuuuuuuuu */
+-
+-/* mm0: 0x000000yy00yy00yy mm1: 0x0000vvvvvvvv0000 mm2: 0x00000000uuuuuuuu */
+-	paddsw %%mm1, %%mm0;        /* Add v to result */
+-	paddsw %%mm2, %%mm0;        /* Add u to result 0x0000bbbbggggrrrr */
+-	psraw $6, %%mm0;           /* Demote precision */
+-	packuswb %%mm0, %%mm0;     /* Pack into RGBA 0x0000000000bbggrr */
+-	movd %%mm0, (%3);          /* Store output */
++asm("\n\
++/* Output will be 0x00bbggrr with the 00 trailing so this can also be used */\n\
++/* for rgb24. */\n\
++	movd (%0), %%mm0;          /* Load y   0x00000000000000yy */\n\
++	psubsw mpeg3_MMX_601_Y_DIFF, %%mm0;      /* Subtract 16 from y */\n\
++	movd (%1), %%mm1;          /* Load v    0x00000000000000vv */\n\
++	movq %%mm0, %%mm3;         /* Copy y to temp */\n\
++	psllq $16, %%mm1;          /* Shift v   0x0000000000vv0000 */\n\
++	movd (%2), %%mm2;          /* Load u    0x00000000000000uu */\n\
++	psllq $16, %%mm3;          /* Shift y */\n\
++	movq %%mm1, %%mm4;         /* Copy v to temp */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x0000000000yy00yy */\n\
++	psllq $16, %%mm4;          /* Shift v */\n\
++	movq %%mm2, %%mm5;         /* Copy u to temp */\n\
++	psllq $16, %%mm3;          /* Shift y  */\n\
++	por %%mm4, %%mm1;          /* Overlay new v byte 0x000000vv00vv0000 */\n\
++	psllq $16, %%mm5;          /* Shift u  */\n\
++	por %%mm3, %%mm0;          /* Overlay new y byte 0x000000yy00yy00yy */\n\
++	por %%mm5, %%mm2;          /* Overlay new u byte 0x0000000000uu00uu */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy     mm1: 0x000000vv00vv0000     mm2: 0x0000000000uu00uu */\n\
++	pmullw mpeg3_MMX_601_Y_COEF, %%mm0;     /* Scale y coeffs */\n\
++ 	psubw mpeg3_MMX_V_80_RGB, %%mm1;    /* Subtract 128 from v 0x000000vv00vv0000 */\n\
++ 	pmullw mpeg3_MMX_V_COEF_RGB, %%mm1; /* Multiply v coeffs 0x0000vvvvvvvv0000 */\n\
++ 	psubw mpeg3_MMX_U_80_RGB, %%mm2;    /* Subtract 128 from u 0x0000000000uu00uu */\n\
++ 	pmullw mpeg3_MMX_U_COEF_RGB, %%mm2; /* Multiply u coeffs 0x0000uuuuuuuuuuuu */\n\
++\n\
++/* mm0: 0x000000yy00yy00yy mm1: 0x0000vvvvvvvv0000 mm2: 0x00000000uuuuuuuu */\n\
++	paddsw %%mm1, %%mm0;        /* Add v to result */\n\
++	paddsw %%mm2, %%mm0;        /* Add u to result 0x0000bbbbggggrrrr */\n\
++	psraw $6, %%mm0;           /* Demote precision */\n\
++	packuswb %%mm0, %%mm0;     /* Pack into RGBA 0x0000000000bbggrr */\n\
++	movd %%mm0, (%3);          /* Store output */\n\
+ 	"
+ :
+ : "r" (&y), "r" (&v), "r" (&u), "r" (output));